Research On Clustering And Resource Allocation Algorithm For Ultra-Dense Network

The Ultra-Dense Network(UDN)is proposed to address the ultra-high capacity and ultra-high user experience rate requirements of 5G and advanced wireless communication system.However,the ultra-dense deployment of small cells also brings a series of problems such as cell association selection,resource allocation,interference management,and mobility management In order to address those problem,the industry proposed the concept of"User-Centric".In the user-centric UDN,multiple small cells around the user are dynamically organized into a cluster.Multiple small cells cooperate to provide data transmission services for an user.This paper focus on the user-centeric clustering and resource allocation in UDN.The main works as follow.Firstly,dynamically constructing small cells cluster for each user in UDNs is very important which will affect the performance of UDN.Therefore,this paper proposes a many-to-many matching based clustering algorithm for the macro-micro-heterogeneous UDN scenario.The algorithm is based on the bilateral many-to-many matching market,consider the constraints of small cells'access capacity restrict and the mutual influence of users'access capacity,under the optimization goal of maximizing system reachability,constructs a stable match between the users and the small cells.Simulation result shows that the proposed small cell clustering algorithm based on many-to-many matching improves the achievable sum rate of the system by nearly 40%compared with the Max-RSRP algorithm,and the users' average achievable rate increases by about 18.75%.Secondly,users access multiple small cells and then face a more complex resource allocation problem.The limited resources and orthogonal constraints make the traditional single-cell resource allocation scheme no longer applicable,so,this paper proposes a resource allocation algorithm based on network decomposition,considers the interference situation in UDN,modeling the heterogeneous network,and designs the optimization objective function of resource allocation.The resource allocation problem is divided into two sub-problems.The network decomposition algorithm maximizes interference within the subnet while minimizing interference between subnets,and implements multiplexing and orthogonal allocation of system resource blocks.Simulation result shows that,compared with the Graph Coloring algoritlhm based on graph coloring theory,the proposed algorithm improves the system throughput by about 10%and the average user spectrum efficiency increased by about 18%,which means the proposed algorithm effectively reduces the interference in the network and improves the user spectrum efficiency.The main contribution of this paper is the design and simulation verification of small cell clustering algorithm and resource allocation algorithm.In the further work,we will focus on reducing the complexity of the algorithm and consider a distributed methed.Besides,joint resource allocation algorithms for other wireless resources such as computing resource and power resource will be further studied.